Hepatic chemoembolization (HCE) routinely results in severe pain requiring massive doses of intravenously administered narcotics. This study examines the efficacy and safety of lidocaine administered intraarterially for analgesia in HCE. In 45 HCE procedures, lidocaine was injected into hepatic arterial branches just prior to and during chemoembolization. Adjunctive analgesic doses given during the procedure and the need for a morphine sulfate drip infusion for postprocedural pain control were recorded and compared with those in 20 procedures performed previously without lidocaine. In procedures with lidocaine, an average of 0.13 mg of morphine sulfate and 1.3 mg of midazolam were required. This is significantly lower than the 11.7 mg of morphine sulfate and 3.7 mg of midazolam used during procedures without lidocaine. A postprocedural morphine drip infusion was required for control of severe pain in 16 of 20 (80%) procedures performed without lidocaine compared with nine of 45 (20%) of those performed with lidocaine. Peripheral blood levels of lidocaine were well below the toxic level, and no complications referable to lidocaine toxicity occurred. Marked reductions in the amount of narcotic analgesia in HCE procedures may be safely achieved with the administration of intraarterial lidocaine.